FRM Part I · FRM Exam Part I · Machine Learning and Prediction
A bank tests a default-prediction classifier on 200 loans. The model flags 40 loans as defaults, of which 30 actually defaulted. In total 50 loans in the sample defaulted. What is the model's precision (positive predictive value)?
Precision is 75%. It is the share of loans the model flagged as defaults that really defaulted: 30 true positives divided by 40 flagged loans. Dividing by the 50 actual defaults would give recall of 60%, which measures a different thing.
- A60%
- B75%Correct
- C80%
- D85%
Explanation
Precision = true positives / predicted positives = 30/40 = 75%. Recall is 30/50 = 60%, which is the common mistake of dividing by actual defaults. Accuracy or other ratios do not equal precision.
Did you get it right without looking?
One question tells you little. A timed set on Machine Learning and Prediction shows your real accuracy, how long you take and where you lose marks.
More Machine Learning and Prediction questions
- A data scientist increases the complexity of a prediction model by adding many more predictors and higher-order terms. Holding the training …
- An analyst has 1,000 monthly observations and uses a 60/20/20 split into training, validation and test sets. Which statement about the use o…
- A risk analyst wants to group a bank's corporate borrowers into clusters based on financial ratios, with no predefined default labels availa…
- A risk analyst builds a model to predict loan defaults. The model achieves 99% accuracy on the training data but only 70% accuracy on a held…
- A PCA on six standardized predictors yields eigenvalues of 3.0, 1.5, 0.6, 0.45, 0.3 and 0.15. What is the smallest number of components need…
- In training a deep neural network, what is the purpose of the backpropagation algorithm?